GBE
Evidence for Adaptation to the Tibetan Plateau Inferred from
Tibetan Loach Transcriptomes
Ying Wang1,2, Liandong Yang1,2, Kun Zhou3, Yanping Zhang4, Zhaobin Song5, and Shunping He1,*
1
The Key Laboratory of Aquatic Biodiversity and Conservation of Chinese Academy of Sciences, Institute of Hydrobiology, Chinese Academy of
Sciences, Wuhan, China
2
University of the Chinese Academy of Sciences, Beijing, China
3
Hubei Key Laboratory of Genetic Regulation and Integrative Biology, College of Life Science, Central China Normal University, Wuhan, China
4
Gansu Key Laboratory of Cold Water Fishes Germplasm Resources and Genetics Breeding, Gansu Fishers Research Institute, Lanzhou, China
5
Sichuan Key Laboratory of Conservation Biology on Endangered Wildlife, College of Life Sciences, Sichuan University, Chengdu, China
*Corresponding author: E-mail: [email protected].
Accepted: October 5, 2015
Data deposition: This project has been deposited at the National Center for Biotechnology Information (NCBI) Sequence Read Archive database
under the accession SRR1946837 for Triplophysa siluroides and SRR1948020 for Triplophysa scleroptera.
Abstract
Triplophysa fishes are the primary component of the fish fauna on the Tibetan Plateau and are well adapted to the high-altitude
environment. Despite the importance of Triplophysa fishes on the plateau, the genetic mechanisms of the adaptations of these fishes
to this high-altitude environment remain poorly understood. In this study, we generated the transcriptome sequences for three
Triplophysa fishes, that is, Triplophysa siluroides, Triplophysa scleroptera, and Triplophysa dalaica, and used these and the previously
available transcriptome and genome sequences from fishes living at low altitudes to identify potential genetic mechanisms for the
high-altitude adaptations in Triplophysa fishes. An analysis of 2,269 orthologous genes among cave fish (Astyanax mexicanus),
zebrafish (Danio rerio), large-scale loach (Paramisgurnus dabryanus), and Triplophysa fishes revealed that each of the terminal
branches of the Triplophysa fishes had a significantly higher ratio of nonsynonymous to synonymous substitutions than that of
the branches of the fishes from low altitudes, which provided consistent evidence for genome-wide rapid evolution in the Triplophysa
genus. Many of the GO (Gene Ontology) categories associated with energy metabolism and hypoxia response exhibited accelerated
evolution in the Triplophysa fishes compared with the large-scale loach. The genes that exhibited signs of positive selection and rapid
evolution in the Triplophysa fishes were also significantly enriched in energy metabolism and hypoxia response categories. Our
analysis identified widespread Triplophysa-specific nonsynonymous mutations in the fast evolving genes and positively selected
genes. Moreover, we detected significant evidence of positive selection in the HIF (hypoxia-inducible factor)-1A and HIF-2B genes
in Triplophysa fishes and found that the Triplophysa-specific nonsynonymous mutations in the HIF-1A and HIF-2B genes were
associated with functional changes. Overall, our study provides new insights into the adaptations and evolution of fishes in the
high-altitude environment of the Tibetan Plateau and complements previous findings on the adaptations of mammals and birds to
high altitudes.
Key words: Tibetan Plateau, adaptation, transcriptome, accelerated evolution, Triplophysa fishes.
Introduction
The cold climate on the Tibetan Plateau, with hypoxic conditions and strong ultraviolet radiation, creates environments
that are extremely hostile to most organisms (Bickler and
Buck 2007; Cheviron and Brumfield 2012). Nevertheless, several animals living in the Tibetan Plateau are well adapted to
these harsh environmental conditions (Gansser 1964).
Determination of the selective and neutral forces that drive
molecular evolution is fundamental to understanding the genetic bases for the adaptations to the high-altitude
environments.
Recent literature has primarily focused on the use of genomic approaches to study the genetic mechanisms of highaltitude adaptations in terrestrial endothermic vertebrates. A
ß The Author(s) 2015. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits noncommercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact [email protected]
2970 Genome Biol. Evol. 7(11):2970–2982. doi:10.1093/gbe/evv192 Advance Access publication October 9, 2015
GBE
Tibetan Loach Transcriptomes
set of candidate genes and the expansions of gene families
related to energy metabolism and hypoxia response were
detected in Tibetan people (Beall et al. 2010; Bigham et al.
2010; Xu et al. 2011), yak (Qiu et al. 2012), Tibetan antelope
(Ge et al. 2013), ground tit (Cai et al. 2013; Qu et al. 2013),
Tibetan mastiff (Gou et al. 2014; Li et al. 2014; Wang et al.
2014), and Tibetan grey wolf (Zhang et al. 2014), demonstrating the genetic mechanisms underlying highland adaptations.
However, to date, only limited amounts of genetic or other
information are available on the adaptations of poikilothermic
species, such as reptiles, amphibians, and fishes, to the cold,
hypoxic conditions of high-altitude habitats. Yang et al. (2012,
2014) compared the transcriptomes of two frogs on the
Tibetan Plateau, from a high-elevation and a low-elevation
site, to identify candidate genes involved in the adaptations
to high elevations, and a similar study was conducted with
lizards. Although these observations provided important
insights into the evolutionary constraints of adaptations to
extreme conditions, our understanding of the genetic bases
of such adaptations in fish remains limited. A comprehensive
transcriptome analysis of a Tibetan fish, Gymnodiptychus
pachycheilus, revealed accelerated genic evolution at the
genomic scale (Yang et al. 2015); however, many species of
Schizothoracinae, to which G. pachycheilus belongs, are polyploid, as reported in previous studies (Zan et al. 1985; Yu et al.
1990). In polyploidy, typically from whole-genome duplication, one copy is free to evolve neutrally and thereby accumulates random mutations more frequently (Ohno 1970;
Selmecki et al. 2015). Because of the polyploidy in these species, whether the accelerated genic evolution of G. pachycheilus is the result of adaptation to the Tibetan Plateau or just the
effect of gene duplications is unclear.
Uplift of the Tibetan Plateau contributed to the speciation
of the Triplophysa fishes (Chen and Zhu 1984). The
Triplophysa fishes are species with distributions at the highest
altitudes of the entire Tibetan Plateau ichthyofauna. The
genus Triplophysa is a strongly diverged fish group, with
137 currently recognized species, most of which occur on
the Tibetan Plateau and adjacent regions (Zhu 1989; Froese
and Pauly 2014). As the primary component of the fish fauna
on the Tibetan Plateau, Triplophysa fishes evolved specific
morphological characteristics for adaptation to the highland
environment, such as gradually disappeared scales (scaleless)
for withstanding cold, abundant blood vessels in gill filament
and pituitary for efficient breathing, black peritoneum for responding to ultraviolet radiation, and secondary sexual characters of male for reproduction (Zhu 1989; Wu Y and Wu C
1992; He et al. 2006; Liu et al. 2009; Ren et al. 2011).
Therefore, these plateau-dwelling loaches are an iconic
symbol for studies of highland adaptations in fishes; however,
the genetic bases for those adaptations remain unknown.
The objective of this study was to identify evolutionary patterns, candidate genes, and lineage-specific nonsynonymous
mutations that might have facilitated adaptations to the
Tibetan Plateau in Triplophysa fishes. In this study, the transcriptomes of Triplophysa siluroides, Triplophysa scleroptera,
and Triplophysa dalaica were determined in our laboratory,
and we used these transcriptomes with the transcriptome of
large-scale loach (Paramisgurnus dabryanus) (Li et al. 2015)
and other previously available genome sequences of zebrafish
(Danio rerio) and cave fish (Astyanax mexicanus) to investigate
changes in evolutionary rates and to identify the potential
genetic bases for high-altitude adaptations of Triplophysa
fishes.
Materials and Methods
Sampling, RNA Extraction, and Sequencing
The ethics committee of the Institute of Hydrobiology,
Chinese Academy of Sciences, approved all of the experimental procedures. Two wild male Triplophysa fishes (T. siluroides
and T. scleroptera) were collected from the upper reach of the
Yellow River in Ruoergai County, Sichuan Province. To obtain
as many expressed genes as possible, five tissues (heart, liver,
brain, spleen, and kidney) were sampled and immediately
placed in liquid nitrogen for storage at 80 C until use.
Total RNA isolation, RNA-Seq library construction, and sequencing were described in our previous study (Wang et al.
2015). All sequence reads have been deposited in the National
Center for Biotechnology Information (NCBI) Sequence Read
Archive database (accession numbers: SRR1946837 for
T. siluroides and SRR1948020 for T. scleroptera).
Data Filtering, De Novo Assembly, and Annotation
The quality of the raw reads was first checked with FastQC
(http://www.bioinformatics.bbsrc.ac.uk/projects/fastqc/), and
then the reads were filtered by removing reads that contained
adapters, reads that contained poly-N and low-quality reads
(<Q20) using the Cutadapt application (version 1.6; http://
code.google.com/p/cutadapt/) and TrimGalore software
(Barbraham, Boinformatics) (http://www.bioinformatics.bab
raham.ac.uk/projects/trim_galore/).
After reads were filtered, the clean Illumina RNA-Seq reads
of T. siluroides and T. scleroptera were de novo assembled
using the Trinity software (Grabherr et al. 2011) with default
parameters. The program CD-HIT-EST was used to remove
redundant transcripts with a sequence identity threshold of
0.95 and a word length of 10 to produce the final assembly
(Li and Godzik 2006). The assembled sequences were annotated by searching for sequence homology against the NCBI
nonredundant protein (NR) database (released on May 3,
2014, at http://www.ncbi.nlm.nih.gov) using BLASTx
(Altschul et al. 1997) with a cutoff E-value set at 1e-5.
Ortholog Identification and Alignment
A set of core-orthologs was constructed from eight teleost
genome species including zebrafish (D. rerio), medaka
Genome Biol. Evol. 7(11):2970–2982. doi:10.1093/gbe/evv192 Advance Access publication October 9, 2015
2971
GBE
Wang et al.
(Oryzias latipes), tetraodon (Tetraodon nigroviridis), fugu
(Takifugu rubripes), stickleback (Gasterosteus aculeatus), cod
(Gadus morhua), platyfish (Xiphophorus maculates), and
tilapia (Oreochromis niloticus). All 25,573 one-to-one coreorthologous annotated proteins were downloaded from
Ensembl (release 70) and were selected as “primer taxa”
which served as the input to generate the core-ortholog database for the program HaMStR v.13.2.3 (Ebersberger et al.
2009) to search for corresponding orthologs in P. dabryanus,
T. scleroptera, T. siluroides, and T. dalaica. Zebrafish was used
as the representative in the core-orthologs database. Ensemblannotated one-to-one orthologous genes of D. rerio and A.
mexicanus were retrieved using Biomart (Vilella et al. 2009).
Large-scale loach (P. dabryanus) is relatively closely related to
Triplophysa fishes, diverged from Triplophysa fishes about 30
Ma (Frickhinger 1991; Nelson 2006). Moreover, zebrafish has
rich genome data, which can be used to assign and characterize the function of the identified adaptively evolutionary
orthologous genes. When more than one corresponding
ortholog was identified, the longest one was chosen using
our in-house scripts. HaMStR v.13.2.3 was run with strict
parameters (-sequence file, -est, -hmmset, -refspec, -representative, and -ublast). Each corresponding orthologous group of
six species was extracted with custom Perl scripts to generate
multiple sequence alignments.
All the orthologous gene sets were aligned at the codon
level with the option “-codon” using the Prank program
(Loytynoja and Goldman 2005) and then trimmed using the
Gblocks program with the parameter “-t=c” (Castresana
2000). The Gblocks program was used to remove potentially
unreliable regions. Additionally, to eliminate the effect of
uncertain bases on the test of positive selection, all positions
that had gaps (“-”) and “N” in the alignments were deleted.
After the trimming process, alignments shorter than 120 bp
were discarded.
Phylogenetic Analyses and Evolutionary Rate Estimation
Concatenated alignments constructed from all orthologs were
used to construct a phylogenetic tree for these six species. The
concatenate data were first partitioned by genes due to the
different evolutionary rates. PartitionFinder software was used
to determine the optimal partition scheme and corresponding
best-fitting nucleotide substitution model under the Bayesian
information criterion (Lanfear et al. 2012). The maximumlikelihood (ML) analysis was implemented in RAxML 7.0.3
(Stamatakis 2006). A total of 1,000 nonparametric bootstrap
replicates were performed with GTRGAMMA substitution
model applied to all partitions. The best-coring ML tree with
branch lengths and bootstrap support values was obtained
and used in the subsequent analyses.
The codeml program in the PAML package (Yang 2007)
with the free ratio model (model = 1) was used to estimate the
evolutionary rate along each lineage for the six species.
The lineage-specific mean Ka/Ks ratios (the ratios of the
number of nonsynonymous substitutions per nonsynonymous
site [Ka] to the number of synonymous substitutions per synonymous site [Ks]) were estimated for each ortholog, the concatenated alignments constructed from all orthologs, and the
1,000 concatenated alignments constructed from ten randomly chosen orthologs. The resulting codeml data (such as
the N, S, dN, dS, and dN/dS values) for the genes of each
lineage were calculated. Genes with N*dN or S*dS < 1 or
dS > 2 were filtered following the approach used in a previous
study (Goodman et al. 2009). Comparisons of the evolutionary rates along each lineage of the six species were conducted
using the Wilcoxon rank sum test.
To identify rapidly evolving GO (Gene Ontology) categories
in Triplophysa fishes relative to P. dabryanus, the average dN/
dS values for each GO category were calculated, with only GO
categories containing more than ten orthologous genes included in these analyses. The Wilcoxon rank sum test was
used to identify GO categories with significantly higher dN/
dS values in Triplophysa fishes or in the P. dabryanus lineage.
Identification of Fast Evolving Genes and Positively
Selected Genes
The branch model was used to identify the fast evolving genes
(FEGs) in the codeml program in the PAML package (Yang
2007). The null model assumed that all branches of the tree
evolved at the same rate (the same o), and the alternative
model allowed for a specifically tested branch (foreground
branch) to evolve under a different rate. A likelihood ratio
test (LRT) (df = 1) was used to discriminate between alternative
models for each ortholog in the gene set. We corrected for
multiple testing by applying the false discovery rate method
(FDR < 0.05) (Benjamini and Hochberg 1995) as implemented
in R (http://www.R-project.org). Genes were considered FEGs
if they had an FDR-adjusted P value < 0.05 and a higher o
value in the foreground branch than in the background
branches.
To find genes that potentially experienced positive selection, the improved branch-site model (model = 2, Nsites = 2) in
the codeml program in the PAML package was used to detect
signatures of positive selection on individual codons in a specific branch (Zhang et al. 2005). By setting the Triplophysa
branches as the foreground branch, we compared a selection
model that allowed a class of codons on that branch to have
dN/dS > 1 (ModelA2, fix_omega = 0, omega = 1.5) with a
neutral model that constrained this additional class of sites
to have dN/dS = 1 (ModelA1, fix_omega = 1, omega = 1). An
LRT compared a model with positive selection on the foreground branch to a null model in which no positive selection
occurred on the foreground branch and calculated the statistic
(2ln) to obtain a P value. As noted above, we performed a
correction for multiple testing using an FDR criterion and
2972 Genome Biol. Evol. 7(11):2970–2982. doi:10.1093/gbe/evv192 Advance Access publication October 9, 2015
GBE
Tibetan Loach Transcriptomes
reported genes with an adjusted P value < 0.05 as positively
selected genes (PSGs).
After FEGs and PSGs were detected, the DAVID Functional
Annotation tool was used to perform GO functional enrichment analyses (Dennis et al. 2003; Huang et al. 2009). Within
each annotation cluster, DAVID listed the GO terms that were
significantly enriched.
Triplophysa Lineage-Specific Nonsynonymous Mutation
Identification and Analysis
Using our in-house scripts, Triplophysa lineage-specific nonsynonymous mutations were identified in amino acid sites of all
orthologous genes, FEGs and PSGs. Using the lineage-specific
nonsynonymous mutations of all the orthologous genes as the
background, the nonsynonymous substitution rates of the
FEGs and the PSGs were compared with that of all orthologous genes with chi-square tests. Additionally, we compared
the numbers of Triplophysa lineage-specific nonsynonymous
substitutions in the FEGs and non-FEGs, in the PSGs and nonPSGs, and in the overlap and nonoverlap of the FEGs and PSGs
and then performed Wilcoxon rank sum tests to determine
the significance of the differences.
To further verify the importance of the lineage-specific nonsynonymous mutations, we also focused on several important
functional genes related to hypoxia, namely, the HIF (hypoxiainducible factor) alpha paralogs (HIF-1A/B and HIF-2A/B).
Orthologs of the HIF alpha paralogs in representative fishes
were retrieved from Ensembl (Cunningham et al. 2015) and
were aligned with MEGA4.0 (Tamura et al. 2007). The codeml
branch-site model (model = 2, Nsites = 2) in the PAML package
(version 4.7) was used to detect positive selection, and positively selected sites (PSSs) were identified using the Bayes
empirical Bayes (BEB) method. Using the software SIFT
(Kumar et al. 2009), the functional consequences of the
Triplophysa lineage-specific nonsynonymous mutations were
predicted, and then the positive selected genes were searched
against the PDB database (Rose et al. 2013) to identify suitable
templates for homology modeling. According to the zebrafish
HIF alpha paralogs reference, the 3D-structural models were
predicted using the I-TASSER software (Roy et al. 2010), visualized with PyMOL v1.5 (Schrödinger, LLC, New York).
Structure-based energy calculation was measured with the
Eris server (Yin et al. 2007) with the flexible backbone
method and pre-relaxation options.
Results
A total of 26,362,264 raw 101-bp paired-end reads of
T. siluroides and 48,836,596 raw 101-bp paired-end reads
of T. scleroptera were generated. After trimming of the adapter sequences and removal of sequences of low quality, the de
novo assemblies of the clean reads produced 58,911 and
119,246 unigenes for T. siluroides and T. scleroptera, respectively. Detailed assembly results are summarized in
supplementary table S1 Supplementary Material online, and
the T. dalaica transcriptome was previously available (Wang
et al. 2015). A total of 31,135 (52.8%) unigenes of T. siluroides and 38,825 unigenes (32.6%) of T. scleroptera had significant matches to currently known proteins in the NR
database. A BLASTx top-hit species distribution of the gene
annotations from the NR database showed the highest homology to the zebrafish. Of the BLASTx top-hits, 31,091
(80.1%) matched zebrafish protein sequences in T. scleroptera, whereas 23,022 (73.9%) matched zebrafish protein sequences in T. siluroides.
Accelerated Evolution of the Triplophysa Lineages
A total of 3,217 1:1:1:1:1:1 putative orthologous genes were
first identified for A. mexicanus, D. rerio, P. dabryanus,
T. siluroides, T. scleroptera, and T. dalaica. After our alignment
treatments and trimming for quality control, 2,269 orthologous genes were advanced to the subsequent evolutionary
analyses.
The phylogeny of the six species was constructed from the
concatenated 2,269 orthologous genes (fig. 1A) with the partitioned ML method. Based on the above constructed phylogenetic tree, the ratio of nonsynonymous to synonymous
substitutions (Ka/Ks) for each ortholog were evaluated in the
different branches at the gene level. The free ratio model
(model = 1) in PAML was used (Yang 2007), which allowed
for a separate Ka/Ks ratio for each branch. Averaged for all
2,269 orthologous genes, all three Triplophysa fish branches
(T. siluroides, T. scleroptera, and T. dalaica) had significantly
higher ratios of nonsynonymous to synonymous substitutions
(Ka/Ks) than the remaining branches of the tree (Wilcoxon
rank sum test, P < 2.2 e-16), which indicated accelerated evolution in the Triplophysa lineages (fig. 1B). Furthermore, we
calculated the Ka/Ks ratio for each branch for a concatenated
alignment of 2,269 orthologs and 1,000 concatenated alignments constructed from ten randomly chosen orthologs,
which also revealed that all the Triplophysa fishes had significantly higher Ka/Ks ratios than the other fish branches
(Wilcoxon rank sum test, P < 2.2 e-16) (fig. 1C and D).
Moreover, we examined the Ka/Ks ratio for each gene and
found a larger number of genes with higher Ka/Ks ratios in the
Triplophysa fishes (T. siluroides, T. scleroptera, and T. dalaica)
than in the P. dabryanus lineage (1,153 vs. 414, 659 vs. 386,
and 792 vs. 410, respectively).
To identify lineage-specific accelerated GO categories, the
mean Ka/Ks ratios for the different GO categories were calculated for each of the Triplophysa fishes and the P. dabryanus
branch. The number of GO categories with higher mean Ka/
Ks ratios in the T. siluroides, T. scleroptera, and T. dalaica lineages was significantly greater than that in the P. dabryanus
lineage (406 vs. 25, 383 vs. 45 and 383 vs. 48, respectively).
The statistical analysis found a total of 188, 156, and 150
GO categories with relatively higher evolutionary rates in the
Genome Biol. Evol. 7(11):2970–2982. doi:10.1093/gbe/evv192 Advance Access publication October 9, 2015
2973
GBE
Wang et al.
FIG. 1.—Phylogenetic tree used in this study (A) and the Ka/Ks ratios for the terminal branches obtained from each ortholog (B), concatenated
alignments constructed from all orthologs (C), and 1,000 concatenated alignments constructed from ten randomly chosen orthologs (D). The blue line in
(A) represents the Triplophysa fishes, which are highlighted in light blue. Tda, T. dalaica; Tsc, T. scleroptera; Tsi, T. siluroides; Pda, P. dabryanus; Dre, D. rerio;
Ame, A. mexicanus.
T. siluroides, T. scleroptera, and T. dalaica lineages, respectively, compared with that for the P. dabryanus lineage
(Wilcoxon rank sum test, P < 0.05). The comparative analysis
between Ka/Ks ratios in the lineage of each of the Triplophysa
fishes and in the P. dabryanus lineage revealed that many GO
categories with significantly higher Ka/Ks ratios in Triplophysa
fishes were associated with hypoxia response and energy metabolism: “ATP binding,” “mitochondrion,” “blood coagulation,” “cellular response to hypoxia,” “response to hypoxia,”
and “oxidation-reduction process” (fig. 2A–C and supplementary tables S2–S4 Supplementary Material online).
Positively Selected Genes and FEGs
To identify genes that might be candidates for Triplophysa
lineage-specific adaptations, we applied eight different
branch-site and branch LRTs: Three terminal branches (each
branch of T. dalaica, T. scleroptera, and T. siluroides); the
2974 Genome Biol. Evol. 7(11):2970–2982. doi:10.1093/gbe/evv192 Advance Access publication October 9, 2015
GBE
Tibetan Loach Transcriptomes
internal branch ancestral to T. dalaica and T. scleroptera; the
whole clade of T. dalaica and T. scleroptera; the internal
branch ancestral to T. dalaica, T. scleroptera, and T. siluroides;
the whole clade of T. dalaica, T. scleroptera, and T. siluroides;
and the P. dabryanus branch. The whole clade represented
both the ancestral branch and the terminal branch for a specific lineage. In total, we identified 278 PSGs and 833 FEGs
from all the Triplophysa branches and 50 PSGs and 112 FEGs
in the P. dabryanus branch after conservatively correcting for
multiple testing (supplementary tables S5 and S6,
Supplementary Material online). To identify genes that
might directly contribute to adaptations to the Tibetan
Plateau, we found a subset (n = 196) of genes that were
both FEGs and PSGs from all the Triplophysa branches, and
among these, identified 12 candidate genes related to the
hypoxia response: ADAR, LSP1, HMOX1A, IL6ST, HDAC3,
ANGPTL3, CAST, EPRS, SMOC1, FOXO4, ADAM8B, and
TNFRSF1B (table 1 and supplementary table S7,
Supplementary Material online).
Functional enrichment analyses showed that the PSGs
(278) and the FEGs (833) from all Triplophysa branches were
significantly enriched in functional categories related to
energy metabolism and to adaptation to hypoxia. These categories are biologically relevant for life at high altitudes. The
ATPase genes (ATPase activity, coupled and ATPase activity)
have a role in providing energy. Additionally, “ATP binding,”
“phosphorylation,” “NADP metabolic process” and “cellular
carbohydrate catabolic process” are related to energy metabolism (table 2 and supplementary table S8, Supplementary
Material online), and the “Oxidoreduction coenzyme” and
“Heat shock protein binding” categories might be involved
in hypoxia adaptation (table 2 and supplementary table S8,
Supplementary Material online).
Widespread Triplophysa Lineage-Specific Nonsynonymous
Mutations
Lineage-specific nonsynonymous mutations can contribute to
lineage-specific adaptations. Among the 2,269 orthologous
genes, 1,958 genes possessed Triplophysa lineage-specific
nonsynonymous mutations, whereas 257 of the 278 PSGs,
794 of the 833 FEGs, and 183 of the 196 overlaps of both
FEGs and PSGs possessed Triplophysa lineage-specific nonsynonymous mutations. Chi-square tests showed that the percentage of genes having Triplophysa lineage-specific
nonsynonymous mutations of functional genes (the FEGs,
the PSGs, and the overlap of both FEGs and PSGs) was significantly higher than that of background (P < 1.91 e-012,
P = 0.004, and P = 0.005, respectively) (fig. 3A). More importantly, in comparisons between the number of Triplophysa
lineage-specific nonsynonymous mutations in FEGs and
non-FEGs, in PSGs and non-PSGs, and in the overlap and
nonoverlap of FEGs and PSGs, the number of Triplophysa lineage-specific nonsynonymous mutations were significantly
higher in the FEGs, the PSGs, and the overlap of FEGs and
PSGs (Wilcoxon rank sum test, P < 2.2 e-16, P < 6.161 e-10,
and P < 5.016 e-11, respectively) (fig. 3B).
Selection analysis of the HIF alpha paralogs revealed that
HIF-1A and HIF-2B genes experienced positive selection
(P < 1.905 e-13 and 1.141 e-02, respectively). In the HIF-1A
genes, PSSs were identified using the BEB method, some of
which were also Triplophysa lineage-specific nonsynonymous
mutations, although no PSSs were identified in the HIF-2B
gene. We examined the nonsynonymous mutations in HIF1A gene in detail and found that some Triplophysa lineagespecific nonsynonymous mutations, including S329N, A407R,
S710P, R743T, and L746Y, were PSSs. Amino acid site L712I
was a Triplophysa lineage-specific nonsynonymous mutation
but did not undergo positive selection. The S329N and L712I
were invariant among all other representative fishes, whereas
the remaining variants were variable in the other representative fishes (fig. 4A). Similarly, in HIF-2B gene, R65Q, Q123P,
H129Q, H418L, and P462S exhibited Triplophysa lineagespecific nonsynonymous mutations (fig. 4B). Further analyses
with more distantly related species confirmed the same patterns (supplementary fig. S1, Supplementary Material online).
The prediction of the functional effects of the variants indicated that S329N, S710P, L712I and L746Y in HIF-1A gene
and R65Q, Q123P and H129Q in HIF-2B gene should be deleterious, whereas the others should be tolerated (table 3). The
suitable templates 4H6J and 1L8C were identified for the
homology modeling of the PAC domain and the C-TAD
domain of HIF-1A, respectively, whereas no suitable templates
were identified for the corresponding domains of HIF-2B.
Using the crystal structure of the PAC and C-TAD domains
of human HIF-1A as templates, our homology modeling suggested that the S329N mutation occurred in the PAC domain,
whereas the S710P, L712I and L746Y mutations occurred in
the C-TAD domain (fig. 4C and D). As these mutations were
located in functional domains of the protein and predicted
functional consequences, the Triplophysa lineage-specific
nonsynonymous mutations might be the primary reasons for
high-altitude hypoxic adaptations in these fish.
Discussion
We sequenced and assembled mixed-tissue transcriptomes
from three Triplophysa fishes, that is, T. siluroides, T. scleroptera, and T. dalaica, in combination with previously available
data to identify the underlying genetic mechanisms for adaptation of fishes to high-altitudes, including evolutionary patterns, adaptive evolutionary genes, and lineage-specific
nonsynonymous mutations.
The determination of how species adapt to extreme environments is a central theme for research in evolutionary biology (Smith and Eyre-Walker 2002). Generally, the signals
of adaptive evolution could be detected by an increased
ratio of nonsynonymous to synonymous substitutions
Genome Biol. Evol. 7(11):2970–2982. doi:10.1093/gbe/evv192 Advance Access publication October 9, 2015
2975
GBE
Wang et al.
FIG. 2.—Mean Ka/Ks ratios for each GO category with more than ten orthologs in each of the Triplophysa fishes and P. dabryanus (A–C, Tda, T. dalaica;
Tsc, T. scleroptera; Tsi, T. siluroides). GO categories with statistically significantly higher Ka/Ks ratios in Triplophysa fish (red) and P. dabryanus (blue) are
highlighted. Points with light red and light blue represent GO categories with higher but statistically not significant Ka/Ks ratios in each of the Triplophysa
fishes and P. dabryanus.
(Bakewell et al. 2007). Our results that Tibetan loaches
showed genome-wide accelerated evolution (higher Ka/Ks
ratio) relative to other low-altitude fishes suggest that
Tibetan loaches may have undergone adaptive evolution
that allows them to cope with their extremely inhospitable
high-altitude environments. In addition, there are many GO
categories involved in the hypoxia response and energy metabolism which were found to have evolved faster in each of
the three Triplophysa fishes than the P. dabryanus lineage.
The above evidence of accelerated evolution provided important insights into the mechanisms of high altitude adaptation
of Triplophysa fishes, which were also observed in endothermic animals, such as yak (Qiu et al. 2012) and poikilothermic
species, such as lizards (Yang et al. 2014) and schizothoracine
fishes (Yang et al. 2015).
It is well-known that the environmental conditions are
extremely adverse in high-altitude habitats. The extreme environments necessitate high energy metabolism and hypoxia
2976 Genome Biol. Evol. 7(11):2970–2982. doi:10.1093/gbe/evv192 Advance Access publication October 9, 2015
GBE
Tibetan Loach Transcriptomes
Table 1
Genes Identified as both FEGs (Branch Model) and PSGs (Branch-Site Model) Involved in the Hypoxia Response in Specific Triplophysa Fishes
Lineages
Gene ID (Gene Name)
ENSDARG00000012389 (ADAR)
ENSDARG00000027310 (LSP1)
Description
Branch along Which PSGs
Was Detected
Branch along Which FEGs
Was Detected
Adenosine deaminase, RNA-specific
Ancestral branch of Tda,
Tsc, and Tsi
Terminal branch of Tda and
whole clade of Tda and
Tsc
Ancestral branch of Tda
and Tsc
Terminal branch of Tda and
whole clade of Tda, Tsc,
and Tsi
Whole clade of Tda, Tsc,
and Tsi
Whole clade of Tda, Tsc,
and Tsi
Lymphocyte-specific protein 1
ENSDARG00000027529 (HMOX1A)
Heme oxygenase (decycling) 1a
ENSDARG00000030498 (IL6ST)
Interleukin 6 signal transducer
ENSDARG00000037514 (HDAC3)
Histone deacetylase 3
Terminal branch of Tsi
Angiopoietin-like 3
Terminal branch of Tda and
whole clade of Tda and
Tsc
ENSDARG00000058693 (CAST)
Calpastatin
ENSDARG00000060494 (EPRS)
Glutamyl-prolyl-tRNA synthetase
ENSDARG00000089188 (SMOC1)
SPARC-related modular calcium
binding 1
Terminal branch of Tda, terminal branch of Tsc, terminal branch of Tsi, and
ancestral branch of Tda
and Tsc
Terminal branch of Tda, terminal branch of Tsc, terminal branch of Tsi, and
ancestral branch of Tda
and Tsc
Terminal branch of Tda
ENSDARG00000055792 (FOXO4)
Forkhead box O4
ENSDARG00000044365 (ANGPTL3)
ENSDARG00000057644 (ADAM8B)
A disintegrin and metalloproteinase domain 8b
ENSDARG00000070165 (TNFRSF1B)
Tumor necrosis factor receptor superfamily, member 1B
Ancestral branch of Tda
and Tsc, whole clade of
Tda and Tsc, and whole
clade of Tda, Tsc and Tsi
Terminal branch of Tda,
whole clade of Tda and
Tsc, and whole clade of
Tda, Tsc and Tsi
Terminal branch of Tda, terminal branch of Tsc, terminal branch of Tsi,
whole clade of Tda and
Tsc, and whole clade of
Tda, Tsc and Tsi
Whole clade of Tda, Tsc,
and Tsi
Terminal branch of Tsi,
whole clade of Tda and
Tsc, and whole clade of
Tda, Tsc and Tsi
Terminal branch of Tsi and
whole clade of Tda, Tsc
and Tsi
Terminal branch of Tda, terminal branch of Tsi,
whole clade of Tda and
Tsc, and whole clade of
Tda, Tsc and Tsi
Whole clade of Tda, Tsc
and Tsi
Terminal branch of Tda, ancestral branch of Tda and
Tsc, whole clade of Tda
and Tsc, and whole clade
of Tda, Tsc and Tsi
Whole clade of Tda and Tsc
and whole clade of Tda,
Tsc and Tsi
Ancestral branch of Tda
and Tsc
Terminal branch of Tda,
whole clade of Tda and
Tsc, and whole clade of
Tda, Tsc and Tsi
Terminal branch of Tsi,
whole clade of Tda and
Tsc, and whole clade of
Tda, Tsc and Tsi
NOTE.—Gene identifier from Ensembl (gene ID from zebrafish) and gene description (Description) are provided. Whole clade represents both the ancestral branch and
the terminal branch for specific lineages. Tda, Triplophysa dalaica; Tsc, Triplophysa scleroptera; Tsi, Triplophysa siluroides.
adaptation in indigenous species living in Tibetan plateau. Our
analyses yielded a list of adaptively evolving genes that
included 278 PSGs and 833 FEGs in the Triplophysa lineages,
which are indeed significantly enriched in functional
categories associated with energy metabolism and hypoxia
response. This observation indicated that the genes are influenced by natural selection during the evolution of Triplophysa
lineages. Notably, the hypoxic environment is a vital factor
Genome Biol. Evol. 7(11):2970–2982. doi:10.1093/gbe/evv192 Advance Access publication October 9, 2015
2977
GBE
Wang et al.
Table 2
GO Enrichment Analysis for FEGs and PSGs in all Triplophysa Fishes
Gene Class
Category/GO ID
GO Term
Gene Number
P Value
Fold Enrichment
FEGs
BP/GO:0006733
BP/GO:0044275
BP/GO:0006796
BP/GO:0006793
BP/GO:0006739
MF/GO:0005524
MF/GO:0031072
BP/GO:0006793
BP/GO:0006796
BP/GO:0006468
BP/GO:0016310
MF/GO:0042623
MF/GO:0016887
MF/GO:0004674
Oxidoreduction coenzyme metabolic process
Cellular carbohydrate catabolic process
Phosphate metabolic process
Phosphorus metabolic process
NADP metabolic process
ATP binding
Heat shock protein binding
Phosphorus metabolic process
Phosphate metabolic process
Protein amino acid phosphorylation
Phosphorylation
ATPase activity, coupled
ATPase activity
Protein serine/threonine kinase activity
5
7
36
36
4
59
7
14
14
10
11
8
8
8
0.0140
0.0232
0.0437
0.0437
0.0081
0.0025
0.0106
0.0268
0.0268
0.0659
0.0691
0.0032
0.0090
0.0969
5.2111
3.1267
1.3748
1.3748
9.0958
1.4580
3.7084
1.9094
1.9094
1.9463
1.8402
4.0984
3.3790
2.0231
PSGs
FIG. 3.—Analysis of Triplophysa lineage-specific nonsynonymous mutations. (A) Percentage of genes having Triplophysa lineage-specific nonsynonymous mutations. “All” represents all the orthologous genes. “Overlap” represents a subset of genes that are both FEGs and PSGs. (B) Number of
Triplophysa lineage-specific nonsynonymous mutations among FEGs and non-FEGs, PSGs and non-PSGs, and overlap and nonoverlap of FEGs and PSGs.
Significant differences are indicated by asterisks, based on chi-square test (A) and Wilcoxon rank-sum test (B), **P < 0.01.
that imposes severe physiological challenges on species living
at high altitude (Storz, Scott, et al. 2010). Nevertheless,
Triplophysa fishes that lived in the Tibetan rivers and lakes
have well adapted to these harsh environmental conditions.
Previous studies of the physiological mechanisms of acclimation to hypoxia have been investigated in hypoxia tolerant
fishes, such as Amazonian fish Prochilodus nigricans (Val
et al. 2015) and the scaleless fish Galaxias maculates (Urbina
and Glover 2012). Hypoxia survival in fish requires a wellcoordinated response to either increase the capacity of O2
uptake and delivery to the tissues through blood adjustments,
especially blood oxygen affinity, or reduce oxygen
consumption through strong metabolic rate depression
(Richards 2011; Urbina and Glover 2012; Val et al. 2015).
These adaptive physiological responses of acclimation to
chronic hypoxia principally result from changes in gene expressions and gene mutations (Semenza 2000; Xiao 2015), which
can alter the capacity of fish to endure hypoxia. In particular,
HIF is a master regulator in the hypoxia signaling pathway,
which largely mediated the regulation and coordination of
changes in gene expression in response to hypoxia in fishes
(Xiao 2015). Recent studies in several fishes have shed light on
the molecular and genetic basis underlying their response to
hypoxia with cDNA microarrays, transcriptome sequencing,
2978 Genome Biol. Evol. 7(11):2970–2982. doi:10.1093/gbe/evv192 Advance Access publication October 9, 2015
Tibetan Loach Transcriptomes
GBE
FIG. 4.—Evolutionary analysis and sequence alignments of Triplophysa lineage-specific nonsynonymous mutations across representative fishes based on
positively selected HIF-1A gene (A) and HIF-2B gene (B). The protein coordinates of HIF-1A and HIF-2B referred to the Ensembl ID ENSDARP00000044281
Genome Biol. Evol. 7(11):2970–2982. doi:10.1093/gbe/evv192 Advance Access publication October 9, 2015
2979
GBE
Wang et al.
Table 3
Triplophysa-Specific Nonsynonymous Mutations in the PSGs HIF-1A
and HIF-2B
Gene Name
HIF-1A
HIF-2B
Amino Acid Variant
SIFT Score
Consequence
S329N
A407R
S710P
L712I
R743T
L746Y
R65Q
Q123P
H129Q
H418L
V420L
D437S
P462S
0.01
0.07
0.00
0.01
0.29
0.00
0.03
0.00
0.00
0.66
0.80
1.00
1.00
Deleterious
Tolerated
Deleterious
Deleterious
Tolerated
Deleterious
Deleterious
Deleterious
Deleterious
Tolerated
Tolerated
Tolerated
Tolerated
and proteomics technologies (Gracey et al. 2001; Chen et al.
2013; Liao et al. 2013; Ao et al. 2015). Through comparative
genomic analyses, we identified 12 candidate genes specifically linked to hypoxia response. For example, HMOX1A
(heme oxygenase (decycling) 1a) is induced by a wide variety
of oxidative stresses (Tomaro et al. 1991), which is cloned in
the crucian carp (Carassius carassius) and plays an important
role in preventing hypoxia-induced cell death (Wang et al.
2008). And HDAC3 (HIF-1a-induced histone deacetylase 3)
is essential in response to hypoxia and serves as an essential
corepressor to repress epithelial gene expression (Wu et al.
2011). Additionally, LSP1 (lymphocyte-specific protein 1) and
CAST (calpastatin) appear to be unregulated in response to
hypoxia (Blomgren et al. 1999; Martin-Rendon et al. 2007).
SMOC1 (SPARC-related modular calcium binding 1) is potentially linked to altered vascular structure (Sherva et al. 2007)
and promotes angiogenesis (Awwad et al. 2015). ADAM8B
has a potential role as a hypoxia-dependent protein in the
pathogenesis and evolution of pancreatic cancer
(Valkovskaya 2008). IL6ST and EPRS are listed in the
HypoxiaDB (a database of hypoxia regulated proteins) and
are associated with hypoxia response (Khurana et al. 2013).
These results indicate that hypoxia-regulated genes which
have undergone adaptive evolution in Tibetan loaches represent an adaptation to the extreme environments on the
Tibetan Plateau.
As previously demonstrated with theoretical models, two
populations living in identical environments have an increased
probability of fixing identical mutation by natural selection (Orr
2005). Therefore, to further explore the implications for
Triplophysa lineage-specific adaptations to high-altitude environments, we focused on lineage-specific nonsynonymous
mutations that occurred in functional genes extensively
(PSGs, FEGs, and the overlap of the two). More recently,
lineage-specific mutations were employed to study convergent evolution in marine mammals. Comparative genomic
analysis in that study revealed 15 of the 44 identical nonsynonymous amino acid substitutions were in genes evolving
under positive selection along the marine mammal lineages,
which had known functional associations with the phenotypic
adaptations to swimming, buoyancy and oxygen store in the
aquatic environment (Foote et al. 2015). This is consistent with
our assumption that lineage-specific mutations are responsible
for adaptation to the environments where they lived.
Additionally, lineage-specific mutations in several important
functional genes, including EPAS1 and HBB, were used to account for adaptation to high-altitude hypoxia in humans and
Tibetan mastiffs (Beall et al. 2010; Gou et al. 2014; Wang et al.
2014). More specifically, these studies revealed that several
HIF-pathway genes experienced positive selection in Tibetan
dog and discovered several nonsynonymous mutations in the
EPAS1 gene, of which G305S occurred in a well-defined protein domain (PAS domain) and was deleterious (Gou et al.
2014; Wang et al. 2014). Likewise, positive selection and nonsynonymous mutation analyses of the HIF alpha paralogs in
the present study also demonstrated that HIF-1A and HIF-2B
were subject to positive selection and had several lineagespecific amino acid variants. Positively selected sites were
detected in HIF-1A, including sites S329N, S710P, L712I, and
L746Y. We tentatively infer that the nonsynonymous mutation in HIF-1A, S329N, might play an important role in the
high-altitude hypoxic adaptation of Tibetan loaches, as this
PSS is not only conserved in all representative vertebrates
but also located in the PAC domain, which was proposed to
contribute to the PAS domain fold. However, the remaining
S710P, L712I, and L746Y mutations in HIF-1A were variants in
the other representative vertebrates and were located in the
C-TAD domain that conferred oxygen-dependent regulation
through the hydroxylation of a conserved asparagine residue.
Additionally, the key S329N, S710P, L712I, and L746Y amino
acid mutations that we identified in HIF-1A were predicted to
be damaging, which was also likely to cause functional change
in HIF-1A. Moreover, several experimental studies revealed
that amino acid variants in the alpha- and/or beta-globin
genes can undoubtedly change Hb-O2 affinity in high-altitude
species, including deer mice (Natarajan et al. 2013; Storz,
Runck, et al. 2010; Storz et al. 2009) and Andean hummingbirds (Projecto-Garcia et al. 2013). Therefore, a deep
FIG. 4.—Continued
and ENSDARP00000074832, respectively. (C) Structure model of the PAC domain. The mutated site S329N in the PAC domain is indicated and was
predicted to decrease the thermodynamic stability of the domain (G = 1.42). (D) Structure model of the C-TAD domain of HIF-1A and three mutated sites
S710P, L712I, and L746Y in the C-TAD domain are marked and are predicted to decrease the thermodynamic stability of the domain (G = 0.96 kcal/mol).
2980 Genome Biol. Evol. 7(11):2970–2982. doi:10.1093/gbe/evv192 Advance Access publication October 9, 2015
GBE
Tibetan Loach Transcriptomes
understanding of the contribution of Triplophysa lineagespecific nonsynonymous mutations to adaptation to the
Tibetan Plateau in Tibetan loaches can only be achieved by
experimental and functional genomics in future.
Conclusions
To the best of our knowledge, this study is the first report to
investigate the genetic mechanisms of adaptation to the conditions on Tibetan Plateau in the Triplophysa fishes at the scale
of the genome. In particular, the detection of accelerated evolution, the identified candidate genes, and the Triplophysaspecific variations related to energy metabolism and hypoxia
response provide evidence for adaptations to the high-altitude
environment in the three Triplophysa fishes. These results provide a meaningful contribution to our understanding of the
molecular mechanisms underlying the adaptations to high altitudes and provide additional resources for future studies on
adaptive evolution in Triplophysa fishes.
Supplementary Material
Supplementary figure S1 and tables S1–S8 are available at
Genome Biology and Evolution online (http://www.gbe.
oxfordjournals.org/).
Acknowledgments
The authors thank two anonymous referees. This work was
supported by the Pilot projects (Grant No.XDB13020100).
Literature Cited
Altschul SF, et al. 1997. Gapped BLAST and PSI-BLAST: a new generation
of protein database search programs. Nucleic Acids Res. 25:3389–
3402.
Ao J, et al. 2015. Genome sequencing of the perciform fish Larimichthys
crocea provides insights into molecular and genetic mechanisms of
stress adaptation. PLoS Genet. 11:e1005118.
Awwad K, et al. 2015. Role of secreted modular calcium-binding protein 1
(SMOC1) in transforming growth factor beta signalling and angiogenesis. Cardiovasc Res. 106:284–294.
Bakewell MA, Shi P, Zhang J. 2007. More genes underwent positive
selection in chimpanzee evolution than in human evolution. Proc
Natl Acad Sci U S A. 104:7489–7494.
Beall CM, et al. 2010. Natural selection on EPAS1 (HIF2alpha) associated
with low hemoglobin concentration in Tibetan highlanders. Proc Natl
Acad Sci U S A. 107:11459–11464.
Benjamini Y, Hochberg Y. 1995. Controlling the false discovery rate: a
practical and powerful approach to multiple testing. J R Stat Soc Ser
B Methodol. 57:289–300.
Bickler PE, Buck LT. 2007. Hypoxia tolerance in reptiles, amphibians,
and fishes: life with variable oxygen availability. Annu Rev Physiol.
69:145–170.
Bigham A, et al. 2010. Identifying signatures of natural selection in Tibetan
and Andean populations using dense genome scan data. PLoS Genet.
6:e1001116.
Blomgren K, et al. 1999. Calpastatin is up-regulated in response to hypoxia
and is a suicide substrate to calpain after neonatal cerebral hypoxiaischemia. J Biol Chem. 274:14046–14052.
Cai Q, et al. 2013. Genome sequence of ground tit Pseudopodoces humilis
and its adaptation to high altitude. Genome Biol. 14:R29.
Castresana J. 2000. Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Mol Biol Evol. 17:
540–552.
Chen J, Zhu S. 1984. Phylogenetic relationships of the subfamilies in the
loach family cobitidae (Pisces). Acta Zootaxonmica Sin. 9:201–207.
Chen K, Cole RB, Rees BB. 2013. Hypoxia-induced changes in the zebrafish (Danio rerio) skeletal muscle proteome. J Proteomics. 78:477–485.
Cheviron ZA, Brumfield RT. 2012. Genomic insights into adaptation to
high-altitude environments. Heredity 108:354–361.
Cunningham F, et al. 2015. Ensembl 2015. Nucleic Acids Res. 43:D662–
D669.
Dennis G Jr, et al. 2003. DAVID: database for annotation, visualization,
and integrated discovery. Genome Biol. 4:P3.
Ebersberger I, Strauss S, von Haeseler A. 2009. HaMStR: profile hidden
Markov model based search for orthologs in ESTs. BMC Evol Biol. 9:157.
Foote AD, et al. 2015. Convergent evolution of the genomes of marine
mammals. Nat Genet. 47:272–275.
Frickhinger KA. 1991. Fossilien atlas fishes [English translation 1995]. Melle
(Germany): Mergus. 1088 pp.
Froese R, Pauly D. 2014. FishBase. World Wide Web electronic publication.
Available from: www.fishbase.org, version (11/2014).
Gansser AE. 1964. Geology of the Himalayas. New York: Wiley
Interscience.
Ge RL, et al. 2013. Draft genome sequence of the Tibetan antelope. Nat
Commun. 4:1858.
Goodman M, et al. 2009. Phylogenomic analyses reveal convergent patterns of adaptive evolution in elephant and human ancestries. Proc
Natl Acad Sci U S A. 106:20824–20829.
Gou X, et al. 2014. Whole-genome sequencing of six dog breeds from
continuous altitudes reveals adaptation to high-altitude hypoxia.
Genome Res. 24:1308–1315.
Grabherr MG, et al. 2011. Full-length transcriptome assembly from RNASeq data without a reference genome. Nat Biotechnol. 29:644–652.
Gracey AY, Troll JV, Somero GN. 2001. Hypoxia-induced gene expression
profiling in the euryoxic fish Gillichthys mirabilis. Proc Natl Acad Sci U S
A. 98:1993–1998.
He DK, Chen YX, Chen YF. 2006. Molecular phylogeny and biogeography
of the genus Triplophysa (Osteichthyes:Nemacheilinae) in the Tibetan
Plateau inferred from cytochrome b DNA sequences. Prog Nat Sci.
16:1395–1404.
Huang DW, Sherman BT, Lempicki RA. 2009. Systematic and integrative
analysis of large gene lists using DAVID bioinformatics resources. Nat
Protoc. 4:44–57.
Khurana P, Sugadev R, Jain J, Singh SB. 2013. HypoxiaDB: a database of
hypoxia-regulated proteins. Database (Oxford) 2013:bat074.
Kumar P, Henikoff S, Ng PC. 2009. Predicting the effects of coding nonsynonymous variants on protein function using the SIFT algorithm. Nat
Protoc. 4:1073–1081.
Lanfear R, Calcott B, Ho SY, Guindon S. 2012. Partitionfinder: combined
selection of partitioning schemes and substitution models for phylogenetic analyses. Mol Biol Evol. 29:1695–1701.
Li C, Ling Q, Ge C, Ye Z, Han X. 2015. Transcriptome characterization and
SSR discovery in large-scale loach Paramisgurnus dabryanus
(Cobitidae, Cypriniformes). Gene 557:201–208.
Li W, Godzik A. 2006. Cd-hit: a fast program for clustering and comparing
large sets of protein or nucleotide sequences. Bioinformatics 22:1658–
1659.
Li Y, et al. 2014. Population variation revealed high-altitude adaptation of
Tibetan mastiffs. Mol Biol Evol. 31:1200–1205.
Liao X, et al. 2013. Transcriptome analysis of crucian carp (Carassius auratus), an important aquaculture and hypoxia-tolerant species. PLoS
One 8:e62308.
Genome Biol. Evol. 7(11):2970–2982. doi:10.1093/gbe/evv192 Advance Access publication October 9, 2015
2981
GBE
Wang et al.
Liu X, He X, Li B, Wang Z. 2009. Triplophysa bleekeri pituitary microstructure. Sichuan J Zool. 28:810–812.
Loytynoja A, Goldman N. 2005. An algorithm for progressive multiple
alignment of sequences with insertions. Proc Natl Acad Sci U S A.
102:10557–10562.
Martin-Rendon E, et al. 2007. Transcriptional profiling of human cord
blood CD133+ and cultured bone marrow mesenchymal stem cells
in response to hypoxia. Stem Cells 25:1003–1012.
Natarajan C, et al. 2013. Epistasis among adaptive mutations in deer
mouse hemoglobin. Science 340:1324–1327.
Nelson JS. 2006. Fishes of the world. New York: John Wiley & Sons.
Ohno S. 1970. Evolution by gene duplication. New York: Springer Verlag.
Orr HA. 2005. The probability of parallel evolution. Evolution 59:216–220.
Projecto-Garcia J, et al. 2013. Repeated elevational transitions in hemoglobin function during the evolution of Andean hummingbirds. Proc
Natl Acad Sci U S A. 110:20669–20674.
Qiu Q, et al. 2012. The yak genome and adaptation to life at high altitude.
Nat Genet. 44:946–949.
Qu Y, et al. 2013. Ground tit genome reveals avian adaptation to living at
high altitudes in the Tibetan Plateau. Nat Commun. 4:2071.
Ren S, Yan Z, Ma H, Wang Z. 2011. Observation of gill structure of
Triplophysa bleekeri. Guizhou Agric Sci. 39:128–130.
Richards JG. 2011. Physiological, behavioral and biochemical adaptations
of intertidal fishes to hypoxia. J Exp Biol. 214:191–199.
Rose PW, et al. 2013. The RCSB Protein Data Bank: new resources for
research and education. Nucleic Acids Res. 41:D475–D482.
Roy A, Kucukural A, Zhang Y. 2010. I-TASSER: a unified platform for
automated protein structure and function prediction. Nat Protoc.
5:725–738.
Selmecki AM, et al. 2015. Polyploidy can drive rapid adaptation in yeast.
Nature 519:349.
Semenza GL. 2000. HIF-1: mediator of physiological and pathophysiological responses to hypoxia. J Appl Physiol (1985). 88:1474–1480.
Sherva R, et al. 2007. A whole genome scan for pulse pressure/stroke
volume ratio in African Americans: the HyperGEN study. Am J
Hypertens. 20:398–402.
Smith NG, Eyre-Walker A. 2002. Adaptive protein evolution in Drosophila.
Nature 415:1022–1024.
Stamatakis A. 2006. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models.
Bioinformatics 22:2688–2690.
Storz JF, et al. 2009. Evolutionary and functional insights into the mechanism underlying high-altitude adaptation of deer mouse hemoglobin.
Proc Natl Acad Sci U S A. 106:14450–14455.
Storz JF, Runck AM, Moriyama H, Weber RE, Fago A. 2010. Genetic differences in hemoglobin function between highland and lowland deer
mice. J Exp Biol. 213:2565–2574.
Storz JF, Scott GR, Cheviron ZA. 2010. Phenotypic plasticity and genetic
adaptation to high-altitude hypoxia in vertebrates. J Exp Biol.
213:4125–4136.
Tamura K, Dudley J, Nei M, Kumar S. 2007. MEGA4: Molecular
Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol
Biol Evol. 24:1596–1599.
Tomaro ML, Frydman J, Frydman RB. 1991. Heme oxygenase induction
by CoCl2, Co-protoporphyrin IX, phenylhydrazine, and diamide: evidence for oxidative stress involvement. Arch Biochem Biophys.
286:610–617.
Urbina MA, Glover CN. 2012. Should I stay or should I go?: physiological,
metabolic and biochemical consequences of voluntary emersion upon
aquatic hypoxia in the scaleless fish Galaxias maculatus. J Comp Phys B
Biochem Syst Environ Physiol. 182:1057–1067.
Val AL, Gomes KRM, de Almeida-Val VMF. 2015. Rapid regulation of
blood parameters under acute hypoxia in the Amazonian fish
Prochilodus nigricans. Comp Biochem Physiol A Mol Integr Physiol.
184:125–131.
Valkovskaya NV. 2008. Hypoxia-dependent expression of ADAM8 in
human pancreatic cancer cell lines. Exp Oncol. 30:129–132.
Vilella AJ, et al. 2009. EnsemblCompara GeneTrees: complete,
duplication-aware phylogenetic trees in vertebrates. Genome Res.
19:327–335.
Wang D, Zhong XP, Qiao ZX, Gui JF. 2008. Inductive transcription and
protective role of fish heme oxygenase-1 under hypoxic stress. J Exp
Biol. 211:2700–2706.
Wang GD, et al. 2014. Genetic convergence in the adaptation of dogs and
humans to the high-altitude environment of the Tibetan Plateau.
Genome Biol Evol. 6:2122–2128.
Wang Y, Yang LD, Wu B, Song ZB, He SP. 2015. Transcriptome analysis of
the plateau fish (Triplophysa dalaica): implications for adaptation to
hypoxia in fishes. Gene 565:211–220.
Wu MZ, et al. 2011. Interplay between HDAC3 and WDR5 is essential
for hypoxia-induced epithelial-mesenchymal transition. Mol Cell.
43:811–822.
Wu Y, Wu C. 1992. The fishes of the Qinghai–Xizang Plateau. Chengdu
(China): Sichuan Science and Technology Press.
Xiao W. 2015. The hypoxia signaling pathway and hypoxic adaptation in
fishes. Sci China Life Sci. 58:148–155.
Xu SH, et al. 2011. A genome-wide search for signals of high-altitude
adaptation in Tibetans. Mol Biol Evol. 28:1003–1011.
Yang L, Wang Y, Zhang Z, He S. 2015. Comprehensive transcriptome
analysis reveals accelerated genic evolution in a Tibet fish,
Gymnodiptychus pachycheilus. Genome Biol Evol. 7:251–261.
Yang W, Qi Y, Fu J. 2014. Exploring the genetic basis of adaptation to high
elevations in reptiles: a comparative transcriptome analysis of two
toad-headed agamas (genus Phrynocephalus). PLoS One 9:e112218.
Yang WZ, Qi Y, Bi K, Fu JZ. 2012. Toward understanding the genetic basis
of adaptation to high-elevation life in poikilothermic species: a comparative transcriptomic analysis of two ranid frogs, Rana chensinensis
and R. kukunoris. BMC Genomics 13:588.
Yang Z. 2007. PAML 4: phylogenetic analysis by maximum likelihood. Mol
Biol Evol. 24:1586–1591.
Yin S, Ding F, Dokholyan NV. 2007. Eris: an automated estimator of protein stability. Nat Methods. 4:466–467.
Yu XY, Li YC, Zhou T. 1990. Karyotype studies of Cyprinid fishes in Chinacomparative study of the karyotypes of 8 species of schizothoracine
fishes. J Wuhan Univ. 2:97–104.
Zan R, Liu W, Song Z. 1985. Tetraploid-hexaploid relationship in schizothoracinae. Acta Genet Sin. 12:137–142.
Zhang JZ, Nielsen R, Yang ZH. 2005. Evaluation of an improved branch-site
likelihood method for detecting positive selection at the molecular
level. Mol Biol Evol. 22:2472–2479.
Zhang W, et al. 2014. Hypoxia adaptations in the grey wolf (Canis lupus
chanco) from Qinghai-Tibet Plateau. PLoS Genet. 10:e1004466.
Zhu SQ. 1989. The loaches of the subfamily Nemacheilinae in China
(Cypriniformes: Cobitidae). Nanjing (China): Jiangsu Scientific
Technology Press..
Associate editor: George Zhang
2982 Genome Biol. Evol. 7(11):2970–2982. doi:10.1093/gbe/evv192 Advance Access publication October 9, 2015